Application Gateway System
The Application Gateway System (AGS) is a distributed server system consisting of multiple heterogeneous servers that appears as a single high performance system on the Internet. It can provide stand-alone services or access back-end capabilities such as databases or repositories.
Clusters of server machines are often used to provide a single logical service on the Internet. This is a cost-effective way to increase response time, throughput and availability of the service. The challenge is to set up the service so that the end-user is transparently connected to one of the servers in the cluster. Some of the commonly used techniques that map the machines in a cluster to the same service name include round robin DNS, packet rewriting, and custom client software. Each technique has its strong and weak points.
As part of a research effort sponsored by the National Science Foundation and the National Library of Medicine, we designed and implemented a new approach based on statistical assignment of processors. The AGS addresses some of the shortcomings of existing solutions by providing a highly scalable and efficient adaptive load balancing mechanism without a central point of failure.
Updated 19 Sep 01